# IMAGE PROCESSOR, TREATMENT SYSTEM, AND IMAGE PROCESSING METHOD

According to some embodiments, an image processor includes a perspective image generator, a perspective image acquirer, a corresponding point acquirer, a first calculator and a display image generator. The first calculator updates the position of the one or more first corresponding point on the first perspective image, based on a difference in position between the first and second perspective images. The display image generator generates a first display image including the first and second perspective images, the one or more first corresponding points updated on the first perspective image, and the one or more second corresponding points on the second perspective image. The perspective image generator changes the first direction into a third direction based on the difference in position between the first and second perspective images, and generates an updated one, viewed in the third direction, of the first perspective image.

**Description**

**CROSS-REFERENCE TO RELATED APPLICATION**

This application is based upon and claims the benefit of priority from Japanese Patent Application No. 2013-227226, filed Oct. 31, 2013; the entire contents of which are incorporated herein by reference.

**FIELD**

Embodiments described herein relate generally to an image processor, a treatment system, and an image processing method.

**BACKGROUND**

At the time of performing a radiotherapy treatment, an object to be treated is positioned at the previously-planned position. There is a method of positioning the object to be treated. For example, corresponding points are designated on a first image captured by radiography at the time of planning a treatment and on a second image captured by radiography at the time of actual treatment, and the object to be treated is positioned based on a difference in position between the corresponding points.

It is difficult to designate the corresponding points when the difference is large in posture of the object between at the planning time and at the treatment time.

**BRIEF DESCRIPTION OF THE DRAWINGS**

**DETAILED DESCRIPTION**

According to some embodiments, an image processor may include, but is not limited to, a perspective image generator, a perspective image acquirer, a corresponding point acquirer, a first calculator, and a display image generator. The perspective image generator generates a first perspective image of a target viewed in a first direction, based on three-dimensional volume data of the target. The perspective image acquirer acquires a second perspective image of the target viewed in a second direction. The corresponding point acquirer acquires one or more pairs of first and second corresponding points respectively on the first and second perspective images. The first calculator updates the position of the one or more first corresponding points on the first perspective image, based on a difference in position between the first and second perspective images. The display image generator generates a first display image. The first display image may include, but is not limited to, the first and second perspective images, the one or more first corresponding points updated on the first perspective image, and the one or more second corresponding points on the second perspective image. The perspective image generator changes the first direction into a third direction based on the difference in position between the first and second perspective images, and generates an updated one, viewed in the third direction, of the first perspective image.

In some cases, the processor may further include, but is not limited to, a second calculator. The second calculator calculates the difference by calculating a difference in position between the three-dimensional volume data and the target at the time the second perspective image is captured before the second perspective image is acquired by the perspective image acquirer, based on the one or more first and second corresponding points of the first and second perspective images.

In some cases, the corresponding point acquirer acquires one or more pairs of third and fourth corresponding points respectively on the first and second perspective images. The second calculator updates the difference in position between the first and second perspective images, based on the one or more pairs of first and second corresponding points and the one or more pairs of third and fourth corresponding points. The first calculator updates the one or more first and third corresponding points, based on the difference updated lastly in position between the first and second perspective images. The display image generator generates an updated one of the first display image, the updated one includes the lastly updated one of the first display image including the first perspective image and the one or more first and third corresponding points, which are updated lastly, and the second perspective image.

In some cases, the second calculator updates the difference, based on one or more undeleted pairs of the first and second corresponding points acquired, when any corresponding points have been deleted.

In some cases, the corresponding point acquirer determines whether a designated corresponding point on one of the first and second perspective images is a modified corresponding point to which the existent one of the one or more first and second corresponding points or an additional corresponding point to be added to the existent ones of the one or more first and second corresponding points. The corresponding point acquirer deletes a closest existent one, to the designated corresponding point, of the one or more first and second corresponding points in a case that the corresponding point acquirer determines that the designated corresponding point is the modified corresponding point.

In some cases, the image processor may further include, but is not limited to, a corresponding point modifier that calculates a similarity between first and second partial images. The first partial image includes the first corresponding point which position has been last updated by the first calculator. The second partial image includes the second corresponding point which is paired with that first corresponding point. The corresponding point modifier detects a region having a highest similarity to the second partial image from the first perspective image, if the similarity calculated is smaller than a predetermined reference value. The corresponding point modifier modifies the position of the first corresponding point based on the region detected.

In some cases, the display image generator modifies, based on the difference, an epipolar line on the first perspective image. The epipolar line is associated with the one or more first corresponding points.

In some cases, the perspective image generator changes the third direction into a fourth direction based on the difference in position between the updated one, viewed in the third direction, of the first perspective image and the second perspective image. The perspective image generator generates a further updated one, viewed in the fourth direction, of the first perspective image.

According to other embodiments, a treatment system may include, but is not limited to, an image processor, an irradiator, a detector and a display system. The irradiator irradiates X-rays to a target. The detector detects X-rays transmitted through the target. The detector makes a second perspective image of the target viewed in a second direction based on the X-rays. The image processor may include, but is not limited to, a perspective image generator, a perspective image acquirer, a corresponding point acquirer, a first calculator, and a display image generator. The perspective image generator generates a first perspective image of the target viewed in a first direction, based on three-dimensional volume data of the target. The perspective image acquirer acquires the second perspective image from the detector. The corresponding point acquirer acquires one or more pairs of first and second corresponding points respectively on the first and second perspective images. The first calculator updates the position of the one or more first corresponding point on the first perspective image, based on a difference in position between the first and second perspective images. The display image generator generates a first display image. The first display image may include, but is not limited to, the first and second perspective images, the one or more first corresponding points updated on the first perspective image, and the one or more second corresponding points on the second perspective image. The perspective image generator changes the first direction into a third direction based on the difference in position between the first and second perspective images, and generates an updated one, viewed in the third direction, of the first perspective image. The display system displays the first display image.

In some cases, the system may further include, but is not limited to, a second calculator. The second calculator calculates the difference by calculating a difference in position between the three-dimensional volume data and the target at the time the second perspective image is captured before the second perspective image is acquired by the perspective image acquirer, based on the one or more first and second corresponding points of the first and second perspective images.

In some cases, the corresponding point acquirer acquires one or more pairs of third and fourth corresponding points respectively on the first and second perspective images. The second calculator updates the difference in position between the first and second perspective images, based on the one or more pairs of first and second corresponding points and the one or more pairs of third and fourth corresponding points. The first calculator updates the one or more first and third corresponding points, based on the difference updated lastly in position between the first and second perspective images. The display image generator generates an updated one of the first display image, the updated one includes the lastly updated one of the first display image including the first perspective image and the one or more first and third corresponding points, which are updated lastly, and the second perspective image.

In some cases, the second calculator updates the difference, based on one or more undeleted pairs of the first and second corresponding points acquired, when any corresponding points have been deleted.

In some cases, the corresponding point acquirer determines whether a designated corresponding point on one of the first and second perspective images is a modified corresponding point to which the existent one of the one or more first and second corresponding points or an additional corresponding point to be added to the existent ones of the one or more first and second corresponding points. The corresponding point acquirer deletes a closest existent one, to the designated corresponding point, of the one or more first and second corresponding points in a case that the corresponding point acquirer determines that the designated corresponding point is the modified corresponding point.

In some cases, the image processor may further include, but is not limited to, a corresponding point modifier that calculates a similarity between first and second partial images. The first partial image includes the first corresponding point which position has been last updated by the first calculator. The second partial image includes the second corresponding point which is paired with that first corresponding point. The corresponding point modifier detects a region having a highest similarity to the second partial image from the first perspective image, if the similarity calculated is smaller than a predetermined reference value. The corresponding point modifier modifies the position of the first corresponding point based on the region detected.

In some cases, the display image generator modifies, based on the difference, an epipolar line on the first perspective image. The epipolar line is associated with the one or more first corresponding points.

In some cases, the perspective image generator changes the third direction into a fourth direction based on the difference in position between the updated one, viewed in the third direction, of the first perspective image and the second perspective image. The perspective image generator generates a further updated one, viewed in the fourth direction, of the first perspective image.

According to still other embodiments, an image processing method may include, but is not limited to, the following processes. A first perspective image of a target viewed in a first direction is generated based on three-dimensional volume data of the target. A second perspective image of the target viewed in a second direction is acquired. One or more pairs of first and second corresponding points respectively on the first and second perspective images are acquired. The position of the one or more first corresponding points on the first perspective image is updated based on a difference in position between the first and second perspective images. A first display image is generated. The first display image includes the first and second perspective images, the one or more first corresponding points updated on the first perspective image, and the one or more second corresponding points on the second perspective image. The first direction is changed into a third direction based on the difference in position between the first and second perspective images. An updated one, viewed in the third direction, of the first perspective image is generated.

In some cases, the image processing method may further include, but is not limited to, the following processes. The difference is calculated by calculating a difference in position between the three-dimensional volume data and the target at the time the second perspective image is captured before the second perspective image is acquired, based on the one or more first and second corresponding points of the first and second perspective images. One or more pairs of third and fourth corresponding points respectively on the first and second perspective images are acquired. The difference in position between the first and second perspective images is updated based on the one or more pairs of first and second corresponding points and the one or more pairs of third and fourth corresponding points. The position of the one or more first and third corresponding points is updated based on the difference updated lastly in position between the first and second perspective images. An updated one of the first display image is generated. The updated one includes the first perspective image and the one or more first and third corresponding points, which are updated lastly, and the second perspective image.

According to furthermore embodiments, an image processor may further include, but is not limited to, a processor, and a memory that stores processor-executable instructions that, when executed by the processor, cause the processor to perform the following acts or operations. The processor calculates a difference in position between first and second perspective images. The first perspective image is of a target viewed in a first direction, from three-dimensional volume data of the target. The second perspective image is of the target viewed in a second direction. The processor updates the position of one or more first corresponding points on the first perspective image, based on the difference. The processor changes the first direction into a third direction based on the difference in position between the first and second perspective images. The processor generates an updated one, viewed in the third direction, of the first perspective image.

**First Embodiment**

Hereinafter, an image processor, a treatment system, and an image processing method according to a first embodiment will be described with reference to drawings. **10** according to the first embodiment. The treatment system **10** may include, but is not limited to, an image processor **100**, a display system **200**, and a radiographic imaging apparatus **400**. The treatment system **10** may further include, but is not limited to, a planning apparatus **300**, a treatment apparatus **500**, and a bed **600**.

The image processor **100** detects the difference in position of an object A to be treated (hereinafter, “target A”), such as a patient, based on images taken by the radiographic imaging apparatus **400** at the time of planning a treatment (hereinafter, “planning time”) and at the time of actually performing the treatment (hereinafter, “treatment time”). The display system **200** displays images taken at the planning time and at the treatment time and receives inputs from a user or an operator, such as a surgeon or a medical doctor. The planning apparatus **300** designate a portion of the target A to be subject to the treatment, based on information concerning the target A obtained before the treatment is performed. The radiographic imaging apparatus **400** captures an image of the target A at the treatment time. The treatment apparatus **500** is used for a treatment on the portion of the target A designated by the planning apparatus **300**. The bed **600** shifts, at the treatment time, the target A based on the positional difference detected by the image processor **100** while maintaining a predetermined posture of the target A.

The planning apparatus **300** makes a treatment plan based on an input received through operations by a user, and information, such as images of the inside of the target A which is to be subject to a radiotherapy, a proton therapy, a particle radiotherapy. Images are data captured using a radiographic imaging apparatus that captures an image of the inside of the target A. The radiographic imaging apparatus may be, but is not limited to, an X-ray imaging apparatus, a computed tomography (CT) apparatus, a magnetic resonance imaging (MRI) apparatus, a positron emission tomography (PET) apparatus, a single photon emission computed tomography (SPECT) apparatus, and the like. The images may be either two-dimensional or three-dimensional. In the first embodiment, descriptions are given with respect to a case of using images collected by an X-ray CT apparatus at the planning time.

The planning apparatus **300** may include, but is not limited to, a database **310**, a display **320**, an operation device **330**, and a controller **340**. The database **310** stores image data of the target A. The stored image data of the target A may be voxel data itself captured by capturing the image of the target A. The stored image data of the target A may also be voxel data as raw data captured by subjecting projected data to a correction process, such as logarithmic conversion, offset correction, sensitivity correction, beam hardening correction, or scattered radiation correction. The stored image data of the target A may also be two-dimensional image data reconstructed from the voxel data. Descriptions are given in the present embodiment with respect to a case where the database **310** stores voxel data.

The display **320** displays perspective images under the control of the controller **340**. Descriptions are given in the present embodiment with respect to a case where a perspective image to be used is a perspective image of the target A viewed in a predetermined direction, which is reconstructed from the voxel data stored in the database **310**, that is, a digitally reconstructed radiograph (DRR). It is preferable that the type of the perspective image displayed by the display **320** corresponds to that of an image captured by the radiographic imaging apparatus **400** that will be explained later. This is because use of mutually-associated images enhances precision of the positioning process and makes user input more convenient. Since the case where the radiographic imaging apparatus **400** is an X-ray imaging apparatus is described in the present embodiment, it is preferable that the display **320** displays a DRR image resulting from simulating an image captured by the X-ray imaging apparatus.

The operation device **330** receives an input through operations by a user and supplies information associated with the operations by the user to the controller **340**. The controller **340** controls each unit included in the planning apparatus **300** based on the information associated with the operations received by the operation device **330**. The controller **340** is an information processing device that includes, for example, a central processing unit (CPU), and performs a process based on a particular program. The controller **340** has the database **310** store information indicating a portion of the target A to be subject to a treatment based on a reconstructed image and information associated with the input received through the operations by the user.

The radiographic imaging apparatus **400** may be, but are not limited to, an X-ray imaging apparatus, a CT apparatus, an MRI apparatus, and the like. The radiographic imaging apparatus **400** is an apparatus that captures an image of the target A. In the first embodiment, a case where the radiographic imaging apparatus **400** is an X-ray imaging apparatus is described. The radiographic imaging apparatus **400** may include, but is not limited to, a controller **410** and first and second image captures. The first image capture includes a first ray irradiator **420** and a first ray detector **440**. The second image capture includes a second ray irradiator **430** and a second ray detector **450**.

The first ray detector **440** detects an X-ray irradiated from the first ray irradiator **420** and penetrating the target A. The first ray detector **440** makes or generates a perspective image of the target A based on the detected X-ray. The first ray detector **440** may be, but is not limited to, a flat panel detector (FPD) that receives the X-ray penetrating the target A and converts the received X-ray into a digital signal. Based on the digital signal, the first ray detector **440** generates a perspective image of the target A.

The second ray detector **450** detects an X-ray irradiated from the second ray irradiator **430** and penetrating the target A. The second ray detector **450** makes or generates a perspective image of the target A based on the detected X-ray. Similar to the first ray detector **440**, the second ray detector **450** includes a flat panel detector (FPD) that receives the X-ray penetrating the target A and converts the received X-ray into a digital signal. Based on the digital signal, the second ray detector **450** generates a perspective image of the target A.

Here, the perspective direction in which the first ray detector **440** views the target A differs from that in which the second ray detector **450** views the target A. For example, the first ray irradiator **420**, the second ray irradiator **430**, the first ray detector **440**, and the second ray detector **450** are disposed so that an image capturing surface for the FPD included in the first ray detector **440** is vertical to an image capturing surface for the FPD included in the second ray detector **450**. Here, each of the first and second ray detectors **440** and **450** may include an image intensifier (I.I.), in lieu of the FPD.

The controller **410** controls each unit of the radiographic imaging apparatus **400**. The controller **410** is an information processing device that includes, for example, a CPU, and performs a process based on a particular program. The controller **410** supplies the image processor **100** with perspective images of the target A generated by the first and second ray detectors **440** and **450**.

The image processor **100** supplies the display system **200** with image data indicating an image generated from the voxel data stored by the database **310** of the planning apparatus **300** and the perspective image received from the radiographic imaging apparatus **400**. Additionally, the image processor **100** calculates a difference D in position of the target A between at the planning time and at the treatment time, based on the image generated from the voxel data, the image received from the radiographic imaging apparatus **400**, and the information received from the display system **200** and concerning the input received through the operations by the user. The image processor **100** supplies the bed **600** with positioning information indicating the calculated difference D. The details of image processing, or configuration of the image processor **100** will be described.

The display system **200** may include, but is not limited to, a display **210** and an operation device **220**. The display **210** receives image data from the image processor **100**. The image data may include, but is not limited to, an image generated based on voxel data, and an image captured by the radiographic imaging apparatus **400**. The display **210** displays the image indicated by the received image data.

The operation device **220** receives an input through operations by the user and supplies the image processor **100** with information associated with the input received through the operations. The operation device **220** receives an input through operations which designates a point on the image displayed on the display **210** and supplies information concerning a coordinate of the designated point to the image processor **100**. The operation device **220** may include, for example, a pointing device such as a keyboard or a mouse, a touch panel, and the like. When the operation device **220** is a touch panel, the operation device **220** may be integrated with the display **210**.

The treatment apparatus **500** is an apparatus to be used for subjecting the target A to radiotherapy, a proton therapy, or a particle radiotherapy. The treatment apparatus **500** may include, but is not limited to, a controller **510** and multiple ray irradiators **520**. The controller **510** controls each unit of the treatment apparatus **500**. The controller **510** is a control processing device that includes a CPU, and performs a process based on a particular program. The controller **510** makes the ray irradiators **520** operable after detecting that the bed **600** shifts the target A based on the difference D detected by the image processor **100**. Each of the ray irradiators **520** irradiates, after being made operable, a radio beam, a proton beam, or a particle beam toward the target A based on control by the user. The ray irradiators **520** are disposed such that the radio beams, the proton beams, or the particle beams irradiated from the ray irradiators **520** intersect at one point (isocenter).

The bed **600** may include, but is not limited to, a stage or a chair to place the target A thereon. The bed **600** moves the bed or the chair to place the target A based on the positioning information received from the image processor **100**. Thus, the portion of the target A predetermined at the planning time is aligned on the isocenter.

**100** according to the first embodiment. As shown in **100** may include, but is not limited to, a perspective image generator **110**, a perspective image acquirer **111**, a corresponding point acquirer **112**, a first calculator **113**, a second calculator **114**, and a display image generator **115**. In the image processor **100** and the treatment system **10** of the first embodiment, the isocenter of the treatment apparatus **500** is set to be an original point of a world coordinate system used for the positioning process. Additionally, the world coordinate system with the isocenter set to the original point is used as a three-dimensional coordinate to be used in the image processor **100** and the treatment system **10**. The isocenter is the point of intersection among the beams irradiated from the ray irradiators **520**.

The perspective image generator **110** generates first and second perspective images IP_{1 }and IP_{2 }based on three-dimensional volume data Vd, first and second parameters C_{1 }and C_{2}, and the difference D of the target A. The three-dimensional volume data Vd is voxel data stored by the database **310** of the planning apparatus **300**.

The first parameter C_{1 }includes information concerning the posture of a virtual camera at the time the first perspective image PI_{1 }is generated, and information concerning conversion of the three-dimensional coordinate into an image coordinate system which is for the first perspective image PI_{1}. The first parameter C_{1 }include at least a camera parameter (external and internal parameters) of the virtual camera at the time three-dimensional volume data Vd of the target A is virtually captured.

The second parameter C_{2 }includes information concerning the posture of a virtual camera at the time the second perspective image PI_{2 }is generated, and information concerning conversion of the three-dimensional coordinate into an image coordinate system for the second perspective image PI_{2}. Similar to the first parameter C_{1}, examples of the second parameter C_{2 }include a camera parameter of the virtual camera.

The first perspective image PI_{1 }is a DRR image constructed based on, for example, the three-dimensional volume data Vd and the first parameter C_{1}. Similarly, the second perspective image PI_{2 }is a DRR image constructed based on, for example, the three-dimensional volume data Vd and the second parameter C_{2}. The difference D of the target A is a value indicating the difference in position of the target A between at the time the three-dimensional volume data is captured and at the time the treatment is performed. The difference D of the target A is output from the first calculator **113** and is input to the perspective image generator **110**.

The perspective image generator **110** shifts the three-dimensional volume data Vd to the position at the treatment time based on the difference D of the target A, and then generates the first and second perspective images PI_{1 }and PI_{2 }using the first and second parameters C_{1 }and C_{2}.

The perspective image acquirer **111** receives a pair of perspective images of the target A generated by the radiographic imaging apparatus **400**. Hereinafter, between the pair of perspective images, the perspective image generated by the first ray detector **440** is referred to as a third perspective image PI_{3}, and the perspective image generated by the second ray detector **450** is referred to as a fourth perspective image PI_{4}. The camera parameter of the first ray detector **440** is referred to as a third parameter C_{3}. The camera parameter of the second ray detector **450** is referred to as a fourth parameter C_{4}. The same type of parameters is used as the first and third parameters C_{1 }and C_{3}. The same type of parameters is used as the second and fourth parameters C_{2 }and C_{4}.

The corresponding point acquirer **112** receives from the operation device **220** of the display system **200**, at least one group of four corresponding points on the respective first to fourth perspective images PI′ to PI_{4}. The corresponding points on the first and second perspective images PI_{1 }and PI_{2 }are defined as a first group of points S_{1}. The corresponding points on the third and fourth perspective images PI_{3 }and PI_{4 }are defined as a second group of points S_{2}.

The first calculator **113** calculates the difference D of the target A based on the first and second groups of points S_{1 }and S_{2}, and the first to fourth parameters C_{1 }to C_{4}. The second calculator **114** performs an update process of shifting, based on the first group of points S_{1}, the first and second parameters C_{1 }and C_{2}, and the difference D of the target A, the coordinate of each corresponding point of the first group of points S_{1 }which is on the first or second perspective image to a position approximating to the position of the associated corresponding point of the second group of points S_{2 }which is on the third or fourth perspective images IP_{3 }and IP_{4}.

The display image generator **115** generates a display image DI to be displayed by the display system **200**, using the first to fourth perspective images PI_{1 }to PI_{4}, and the first and second groups of points S_{1 }and S_{2}. The display image generator **115** supplies the display system **200** with image data indicating the generated display image DI. For example, on the display image DI, the first and third perspective images PI_{1 }and PI_{3 }are arranged adjacent to each other, and the second and fourth perspective images PI_{2 }and PI_{4 }are arranged adjacent to each other, thereby enabling easy comparison of the images captured at the planning time and at the treatment time. Here, the corresponding points designated by the user (first and second groups of points S_{1 }and S_{2}) may be superimposed on the first to fourth perspective images PI_{1 }to PI_{4}.

**100** according to the first embodiment. When the image processor **100** initiates the positioning process, the perspective image generator **110** generates the first and second perspective images PI′ and PI_{2 }based on the three-dimensional volume data Vd, the first and second parameters C_{1 }and C_{2}, and the difference D of the target A (step S**101**).

The perspective image acquirer **111** acquires the third and fourth perspective images PI_{3 }and PI_{4 }captured by the radiographic imaging apparatus **400** capturing images of the target A (step S**102**).

The display image generator **115** generates a first display image DI_{1 }including the first and second perspective images PI_{1 }and PI_{2 }generated in step S**101** and the third and fourth perspective images PI_{3 }and PI_{4 }captured in step S**102**. The display image generator **115** supplies the display system **200** with the generated first display image DI_{1}, and has the display system **200** display the first display image DI_{1}, thereby having the user compare the first to fourth perspective images PI_{1 }to PI_{4 }(step S**103**).

The corresponding point acquirer **112** acquires at least one group of four corresponding points (first and second groups of points S_{1 }and S_{2}) designated by the user respectively on the first to fourth perspective images PI_{1 }to PI_{4 }displayed by the display system **200** (step S**104**). The first calculator **113** calculates the difference D of the target A based on the corresponding points captured in step S**104** and the first to fourth parameters C to C_{4 }(step S**105**).

The perspective image generator **110** regenerates first and second perspective images PI_{1}′ and PI_{2}′ based on the difference D of the target A calculated in step S**105**, the three-dimensional volume data Vd, and the first and second parameters C_{1 }and C_{2 }(step S**106**). The second calculator **114** updates the coordinates of the first group of points S_{1 }using the difference D of the target A calculated in step S**105** and the first and second parameters C_{1 }and C_{2 }(step S**107**).

The display image generator **115** generates a second display image DI_{2 }including the first and second perspective images PI_{1}′ and PI_{2}′ regenerated in step S**106**, the first group of points S_{1 }updated in step S**107**, the third and fourth perspective images PI_{3 }and PI_{4 }captured in step S**102**, and the second group of points S_{2}. The display image generator **115** supplies the display system **200** with image data indicating the generated second display image DI_{2}, and has the display system **200** display the second display image DI_{2}, thereby having the user compare the first to fourth perspective images PI_{1}′, PI_{2}′, PI_{3}, and PI_{4}, and the first and second groups of points S_{1 }and S_{2 }(step S**108**).

The corresponding point acquirer **112** determines whether or not an operation by the user, which indicates that input of corresponding points has been complete, is received via the operation device **220** (step S**109**).

If the input of corresponding points has not been complete (step S**109**: NO), the corresponding point acquirer **112** returns the processing to step S**104**.

If the input of corresponding points has been complete (step S**109**: YES), on the other hand, the first calculator **113** supplies the difference D of the target A to the bed **600** and has the bed **600** perform a positioning process (step S**110**), and then terminates the positioning process.

Hereinafter, the above process in each step will be described in detail. **101** and S**106** to be performed by the perspective image generator **110** in the positioning process of the first embodiment. The perspective image generator **110** receives the difference D of the target A from the first calculator **113** (step S**121**). In step **101**, the difference D of the target A has not yet been calculated by the first calculator **113**. For this reason, in step S**101**, a default value predetermined or set by the user may be used as the difference D of the target A.

The difference D of the target A is a value based on rotation and translation of the target A performed to shift the position of the target A at the treatment time to the position of the target A at the time the three-dimensional volume data Vd is captured. For example, the difference D of the target A can be expressed as a 3×3 rotation matrix R and a translation vector t, as shown in the following Expressions (1-1) and (1-2).

Additionally, as the rotation-based value, (ΔR_{x}, ΔR_{y}, ΔR_{z}), which are the rotation angles around the X-axis, the Y-axis, and the Z-axis of the world coordinate system converted from the rotation matrix R, may be used, in lieu of the rotation matrix R. Assuming that rotation is made around the X-axis, the Y-axis, and the Z-axis in this order, the conversion from the rotation matrix R into the rotation angles (ΔR_{x}, ΔR_{y}, ΔR_{z}) can be made by the following Expression (2).

The perspective image generator **110** converts the position of the three-dimensional volume data Vd of the target A into the position of the target A at the treatment time based on the rotation matrix R and the translation vector t which indicate the difference D of the target A (step S**122**). Here, conversion from a position X into a position Y can be made by the following Expression (3) using the rotation matrix R and the translation vector t, where the position X represents the position of the target A at the treatment time, and the position Y represents the position of the three-dimensional volume data Vd.

*V=RX+t* (3)

The position X is obtained from Expression (3), thereby making it possible to shift the position Y of the three-dimensional volume data to a position Y′ which is closer to the position X of the target A at the treatment time. The position Y′ is obtained from the following Expression (4). When the rotation matrix R and the translation vector t are accurate, the position X matches the position Y′.

*Y′=R*^{−1}(*Y−t*) (4)

The perspective image generator **110** generates a first perspective image PI_{1 }by having a first virtual camera VC_{1 }set with the first parameter capture an image of the shifted three-dimensional volume data Vd. Additionally, the perspective image generator **110** generates a second perspective image PI_{2 }by having a second virtual camera VC_{2 }set with the second parameter capture an image of the shifted three-dimensional volume data Vd (step S**123**).

Thus, the perspective image generator **110** generates the first and second perspective images PI_{1 }and PI_{2}.

Hereinafter, procedure for the first calculator **113** to calculate the difference D of the target A in step S**105** is explained. The difference D calculated by the first calculator **113** is indicated by the rotation matrix R and the translation vector t, as explained above. Regarding the n-th group of corresponding points (n≧1) among the N groups of corresponding points captured in step S**104**, a coordinate of a first point of the n-th group of corresponding points, which is on the first perspective image PI_{1 }is represented as (u_{n}^{(1)}, v_{n}^{(1)}). A coordinate of a second point of the n-th group of corresponding points, which is on the second perspective image PI_{2 }and corresponds to the first point (u_{n}^{(1)}, v_{n}^{(1)}), is represented as (u_{n}^{(2)}, v_{n}^{(2)}). A coordinate of a third point of the n-th group of corresponding points, which is on the third perspective image PI_{3 }and corresponds to the first point (u_{n}^{(1)}, v_{n}^{(1)}), is represented as (u_{n}^{(3)}, v_{n}^{(3)}). A coordinate of a fourth point of the n-th group of corresponding points, which is on the fourth perspective image PI_{4 }and corresponds to the first point (u_{n}^{(1)}, v_{n}^{(1)}) is represented as (u_{n}^{(4)}, v_{n}^{(4)}). Additionally, the first group of points S_{1 }is represented as {(u_{n}^{(3)}, v_{n}^{(3)}), (u_{n}^{(4)}, v_{n}^{(4)})} (n=1, . . . , N). The second group of points S_{2 }is represented as {(u_{n}^{(3)}, v_{n}^{(3)}), (u_{n}^{(4)}, v_{n}^{(4)})} (n=1, . . . , N).

The first calculator **113** selects a method of calculating the difference D in accordance with the number of groups of corresponding points N captured by the corresponding point acquirer **112**. There are four different methods of calculating the difference D. The first method is used when N=1. The second method is used when N=2. The third method is used when N=3. The fourth method is used when N≧4. Hereinafter, the four methods of calculating the difference D are sequentially described.

**(First Calculating Method for N=1)**

A first point of the first group of points S_{1}, which is on the first perspective image PI_{1}, is represented as (u^{(1)}, v^{(1)}). A second point of the first group of points S_{1}, which is on the second perspective image PI_{2}, is represented as (u^{(2)}, v^{(2)}). A third point of the second group of points S_{2}, which is on the third perspective image PI_{3}, is represented as (u^{(3)}, v^{(3)}). A fourth point of the second group of points S_{2}, which is on the fourth perspective image PI_{4}, is represented as (u^{(4)}, v^{(4)}). Additionally, the first to fourth parameters C_{1 }to C_{4 }are defined by the following Expression (5).

In Expression (5), j (j=1, 2, 3, 4) represents the parameter number. The three-dimensional coordinate Y=(y_{1}, y_{2}, y_{3}) is calculated from the following Expression (6) using the first point (u^{(1)}, v^{(1)}) on the first perspective image PI_{1}, the second point (u^{(2)}, v^{(2)}) on the second perspective image PI_{2}, and the first and second parameters C_{1 }and C_{2}.

Similarly, the three-dimensional coordinate X=(x_{1}, x_{2}, x_{3}) is calculated using the third point (u^{(3)}, v^{(3)}) of the second group of points S_{2 }which is on the third perspective image PI_{3}, the fourth point (u^{(4)}, v^{(4)}) of the second group of points S_{2}, and the third and fourth parameters C_{3 }and C_{4}. Using the calculated three-dimensional coordinates X and Y, the translation vector t is calculated by the following Expression (7).

At this time, the rotation matrix R is assumed to be a 3×3 unit matrix.

Thus, when one group of corresponding points is captured, the first calculator **113** calculates the rotation matrix R and the translation vector t by the above first calculating method.

**(Second Calculating Method for N=2)**

The first calculator **113** calculates two three-dimensional coordinates Y_{1}=(y(1)_{1}, y(1)_{2}, y(1)_{3}) and Y_{2}=(y(2)_{1}, y(2)_{2}, y(2)_{3}) from two groups of corresponding points included in the first group of points S_{1}. Additionally, the first calculator **113** calculates two three-dimensional coordinates X_{1}=(x(1)_{1}, x(1)_{2}, x(1)_{3}) and X_{2}=(x(2)_{1}, x(2)_{2}, x(2)_{3}) from two groups of corresponding points included in the second group of points S_{2}. Here, calculation of the three-dimensional coordinates is performed in a similar manner as in the case of N=1.

From Expression (3), the three-dimensional coordinates X_{1 }and Y_{1 }meet Y_{1}=RX_{1}+t, and the three-dimensional coordinates X_{2 }and Y_{2 }meet Y_{2}=RX_{2}+t. A vector (Y_{2}−Y_{1}) extending from the three-dimensional coordinate Y_{1 }to the three-dimensional coordinate Y_{2 }can be obtained by the following Expression (8).

It can be understood from Expression (8) that the relationship between the vector (Y_{2}−Y_{1}) extending from the three-dimensional coordinate Y_{1 }to the three-dimensional coordinate Y_{2 }and the vector (X_{2}−X_{1}) extending from the three-dimensional coordinate X_{1 }to the three-dimensional coordinate X_{2 }is dependent only on the rotation matrix R. Here, angle θ between vectors a and b can be expressed by the following Expression (9), where the vector a represents the vector (X_{2}−X_{1}) extending from the three-dimensional coordinate X_{1 }to the three-dimensional coordinate X_{2}, and a vector b represents vector (Y_{2}−Y_{1}) extending from the three-dimensional coordinate Y_{1 }to the three-dimensional coordinate Y_{2}.

In Expression (9), “a·b” represents a scalar product of the vectors a and b. ∥a∥ represents the magnitude of the vector a, and ∥b∥ represents the magnitude of the vector b. Additionally, a unit vector d in the direction of the axis around which the rotation is made by the angle θ is calculated by the following Expression (10).

In Expression (10), “a×b” represents a cross product of the vectors a and b. The vector d represents a vector obtained by normalizing to 1, the magnitude of the vector c obtained from the cross product a×b. The vector b represents a vector obtained by rotating the vector a by the rotation angle θ around the unit vector d. Therefore, the rotation matrix R is calculated by the following Expression (11) using the Rodrigues's formula.

*R*=cos θ*I*+(1−cos θ)*dd*^{T}+sin θ[*d]*_{x} (11)

In Expression (11), I represents a 3×3 unit matrix. [d]_{x }is a 3×3 unit matrix expressed by the Expression (12).

After the rotation matrix R is calculated, the translation vector t is calculated using the following Expression (13) derived from Expression (3).

*t=Y*_{n}*−RX*_{n} (13)

In Expression (13), n may be any one of 1 and 2. Additionally, an average of a translation vector calculated when n=1 and a translation vector calculated when n=2 may be set to be the translation vector t.

Thus, when two groups of corresponding points are captured, the first calculator **113** calculates the rotation matrix R and the translation vector t by the above second calculating method.

**[Third Calculating Method for N=3]**

**113** to calculate the rotation matrix R and the translation vector t when there are three groups of corresponding points in the first embodiment. The first calculator **113** calculates three three-dimensional coordinates Y_{n }(n=1, 2, 3) respectively from three groups of corresponding points included in the first group of points S_{1}, and calculates three three-dimensional coordinates X_{n }(n=1, 2, 3) respectively from the three associated groups of the corresponding points included in the second group of points (step S**131**).

The first calculator **113** calculates the fourth three-dimensional coordinate X_{4 }included in the first group of point S_{1 }using the three-dimensional coordinates Y_{n }calculated in step S**131**, and calculates the fourth three-dimensional coordinate Y_{4 }included in the second group of point S_{2 }using the three-dimensional coordinates X_{n }calculated in step S**131** (step S**132**).

To calculate the fourth three-dimensional coordinate, a gravity center X_{c}=(X_{1}+X_{2}+X_{3})/3 of the three-dimensional coordinates X_{n }(n=1, 2, 3) is calculated. Here, a vector extending from the gravity center X_{c }to the three-dimensional coordinate X_{1 }is represented as a vector e, and a vector extending from the gravity center X_{c }to the three-dimensional coordinate X_{2 }is represented as a vector f. The fourth three-dimensional coordinate X_{4 }is defined as a three-dimensional coordinate obtained by shifting the gravity center X^{c }by an arbitrary length l in the direction of a cross product e×f of the vector e and the vector f.

Similarly, a gravity center Y_{c}=(Y_{1}+Y_{2}+Y_{3})/3 of the three-dimensional coordinates Y_{n }(n=1, 2, 3) is calculated. Then, the fourth three-dimensional coordinate Y_{4 }is calculated using the gravity center Y_{c }and the three-dimensional coordinates Y_{1 }and Y_{2}.

The first calculator **113** calculates a gravity center of the first to fourth three-dimensional coordinates X_{1 }to X_{4 }and a gravity center of the first to fourth three-dimensional coordinates Y_{1 }to Y_{4}, using the following Expressions (14-1) and (14-2) (step S**133**).

In Expression (14-1), “ _{n }(n=1, 2, 3, 4). In

Expression (14-2), “ _{n }(n=1, 2, 3, 4).

The first calculator **113** calculates difference D (indicated by the rotation matrix R and the translation vector t) using the three-dimensional coordinates X_{n }and Y_{n }(n=1, 2, 3), and the gravity centers **134**).

Specifically, each of the three-dimensional coordinates X_{n }(n=1, 2, 3) is shifted by subtracting therefrom the gravity center _{n}′=X_{n}− _{n }is shifted by subtracting therefrom the gravity center _{n}′=− _{1}′, X_{2}′, X_{3}′], and a matrix Y′ is defined as Y′=[Y_{1}′, Y_{2}′, Y_{3}′].

As a result of shifting each of the three-dimensional coordinates by subtracting therefrom the gravity center, the translation vector t is cancelled. For this reason, the relationship between the matrices X′ and Y′ can be expressed, using the rotation matrix R, by the following Expression (15).

*Y′=RX′* (15)

The rotation matrix R meeting Expression (15) can be calculated from Expression (16), using the singular value decomposition (SVD).

In Expression (16), the matrixes U and V are matrixes obtained from the singular value decomposition. Then, the translation vector t is calculated by the following Expression (17), using the rotation matrix R.

*t= X−R Y* (17)

Here, the translation vector t may be calculated using the following Expression (18), in lieu of Expression (17). Alternatively, an average of the translation vectors for n=1, 2, 3, 4 may be set to be the translation vector t.

*t=X*_{n}*−RY*_{n} (18)

Thus, in a case where three groups of corresponding points are captured, the first calculator **113** calculates the rotation matrix R and the translation vector t by the above third calculating method.

**[Calculating Method for N≧4]**

When there are four or more groups of corresponding points, four or more three-dimensional coordinates Y_{n }are obtained from the first group of points S_{1}, and four or more three-dimensional coordinates X_{n }are obtained from the second group of points S_{2}. When N≧4, compared with when N=3, a fourth three-dimensional coordinate needs not be calculated. For this reason, step S**132** in the procedure shown in **113** sequentially performs the processes in steps S**131**, S**133**, and S**134** shown in

Then, procedure for the second calculator **114** to update the coordinates of the first group of points S_{1 }in step S**107** is described. Among the points included in the first group of points S_{1}, the coordinates of the points on the first perspective image PI_{1 }are represented as (u(n)^{(1)}, v(n)^{(1)}), and the coordinates of the points on the second perspective image PI_{2 }are represented as (u(n)^{(2)}, v(n)^{(2)}), where n=1, . . . , N, and N represents the number of groups of corresponding points.

**114** to update the coordinates of the first group of points S_{1 }according to the first embodiment. The second calculator **114** calculates three-dimensional coordinates Y_{n }of the respective points included in the first group of points S_{1}, using Expression (6) (step S**141**).

The second calculator **114** shifts the three-dimensional coordinate Y_{n }using the difference D (indicated by the rotation matrix R and the translation vector t) calculated by the first calculator **113**, thus calculating a three-dimensional coordinate {tilde over (Y)}_{n }(step S**142**). The three-dimensional coordinates {tilde over (Y)}_{n }are positions close to the three-dimensional coordinates X_{n }of the corresponding points included in the second group of points S_{2}. Here, when the rotation matrix R and the translation vector t are accurate, the three-dimensional coordinates {tilde over (Y)}_{n }substantially match the respective three-dimensional coordinates X_{n}.

The three-dimensional coordinates {tilde over (Y)}_{n }are calculated by the following Expression (19), using Expression (4) and the difference D (indicated by the rotation matrix R and the translation vector t).

*{tilde over (Y)}*_{n}*=R*^{−1}(*Y*_{n}*−t*) (19)

The second calculator **114** calculates updated coordinates (u′(n)^{(1)}, v′(n)^{(1)}) on the first perspective image PI_{1 }and updated coordinates (u′(n)^{(2)}, v′(n)^{(2)}) on the second perspective image PI_{2}, which are projected from the three-dimensional coordinate {tilde over (Y)}_{n }using the first and second parameters P_{1 }and P_{2 }defined by Expression (5) (step S**143**). The coordinates (u′(n)^{(1)}, v′(n)^{(1)}) and (u′(n)^{(2)}, v′(n)^{(2)}) of the updated points are calculated using the following Expression (20).

In Expression (20), λ represents the element on the third row of the 3×1 matrix obtained from C_{j}{tilde over (Y)}_{n}.

Thus, the second calculator **114** updates the coordinates of the first group of points S_{1}.

In the aforementioned positioning process, the image processor **100** according to the first embodiment calculates the difference D based on the groups of corresponding points input by the user. Based on the calculated difference D, the image processor **100** regenerates the first and second perspective images PI_{1}′ and PI_{2}′ and updates the coordinates of the groups of corresponding points. The image processor **100** has the display system **200** display a display image DI including the regenerated first and second perspective images PI_{1}′ to PI_{4}′, the third and fourth perspective images PI_{3 }and PI_{4}, and the updated corresponding points. While the regeneration and displaying of the first and second perspective images PI_{1}′ and PI_{2}′, the user can input corresponding points on the perspective images at the planning time (first and second perspective images PI_{1}′ and PI_{2}′) and the perspective images at the treatment time (third and fourth perspective images PI_{3 }and PI_{4}). In a case that there is a difference in posture of the target A between at the planning time and at the treatment time, it is possible to make the perspective images at the planning time more similar to the perspective images at the treatment time, thereby making it easier for the user or operator to designate the corresponding points. In another case that there is another difference in position of the target A between at the planning time and A at the time an X-ray photography is taken, it is also possible to make the perspective images at the planning time more similar to the perspective images at the time the X-ray photography is taken, thereby making it easier for the user or operator to designate the corresponding points. Consequently, it is possible to improve the precision of the positioning process, thereby enabling more precise treatment.

**100** according to the first embodiment. The image processor **100** generates first and second perspective images PI_{1 }and PI_{2 }based on the three-dimensional volume data Vd stored in the database **310** of the planning apparatus **310**. The image processor **100** captures third and fourth perspective images PI_{3 }and PI_{4 }captured by the radiographic imaging apparatus **400** at the treatment time. The image processor **100** supplies the display system **200** with a display image DI including the generated first and second perspective images PI_{1 }and PI_{2 }and the captured third and fourth perspective images PI_{3 }and PI_{4}, thus presenting the display image DI to the user. The image processor **100** receives, through operations by the user, an input of information indicating at least one group of corresponding points indicating the four corresponding portions respectively on the first to fourth perspective images PI_{1 }to PI_{4}. Then, the image processor **100** calculates a difference D indicating the difference in position of the target A between at the planning time and at the treatment time. Based on the difference D, the image processor **100** shifts the three-dimensional volume data Vd and then regenerates first and second perspective images PI_{1}′ and PI_{2}′.

_{2}′ performed by the image processor **100** according to the first embodiment. Specifically, _{1 }to PI_{4 }when at least one group of corresponding points are input. _{1}′, PI_{2}′, PI_{3}, and PI_{4 }after the corresponding points are input. As shown in _{1}′ and PI_{2}′ regenerated from the three-dimensional volume data Vd shifted based on the calculated difference D is changed in such a manner as to reflect the posture of the target A on the third and fourth perspective images PI_{3 }and PI_{4}.

The image processor **100** regenerates the first and second perspective images PI_{1}′ and PI_{2}′ from the three-dimensional volume data Vd based on the calculated difference D and presents the regenerated perspective images PI_{1}′ and PI_{2}′ to the user. Thus, it is possible to present to the user, the first and second perspective images PI_{1}′ and PI_{2}′ respectively similar to the third and fourth perspective images PI_{3 }and PI_{4}. Additionally, corresponding points can be designated on the regenerated first and second perspective images PI_{1}′ and PI_{2}′ respectively similar to the third and fourth perspective images PI_{3 }and PI_{4}, thereby making it easier for the user to designate the corresponding points.

Here, the display image generator **115** of the first embodiment may overlay-display on the first to fourth perspective images PI_{1}′, PI_{2}′, PI_{3}, and PI_{4}, an epipolar line L corresponding to the corresponding points (first and second groups of points S_{1 }and S_{2}) input by the user, as shown in _{1 }and PI_{2 }may be modified in accordance with the difference D.

Additionally, the display image generator **115** of the first embodiment may include combined images I_{1 }and I_{2 }in the display image DI including the first to fourth perspective images PI_{1 }to PI_{4}. The combined image I_{1 }is obtained by combining the first and third perspective images PI_{1 }and PI_{3}, thus superimposing on the combined image I_{1}, the corresponding points of the first group of points S_{1 }on the first perspective image PI_{1 }and the corresponding points of the second group of points S_{2 }on the third perspective image PI_{3}. The combined image I_{2 }is obtained by combining the second and fourth perspective images PI_{2 }and PI_{4}, thus superimposing on the combined image I_{2}, the corresponding points of the first group of points S_{1 }on the second perspective image PI_{2 }and the corresponding points of the second group of points S_{2 }on the fourth perspective image PI_{4}.

_{1 }and I_{2 }according to the first embodiment. As shown in _{1 }and I_{2 }are included in the display image DI such that the combined image I_{1}, the first perspective image PI_{1}′, and the third perspective image PI_{3 }are aligned, and the combined image I_{2}, the first perspective image PI_{2}′, and the third perspective image PI_{4 }are aligned. Thus, the combined images I_{1 }and I_{2 }are displayed with the perspective images PI_{1}′, PI_{2}′, PI_{3}, and PI_{4}, thereby making it possible for the user to recognize the difference in position of the target A between at the planning time and at the treatment time. Thus, it is possible for the user to easily designate corresponding points.

Regarding the positioning process performed by the image processor **100** of the first embodiment, descriptions have been given with respect to the configuration such that it is determined whether or not input of corresponding points by the user has been complete (step S**109**), thereby repeatedly performing calculation of the difference D, regeneration of the first and second perspective images PI_{1}′ and PI_{2}′, an update of coordinates of the first group of points S_{1}. However, the configuration is not limited thereto. For example, the configuration may be such that the calculation of the difference D, the regeneration of the first and second perspective images PI_{1}′ and PI_{2}′, the update of coordinates of the first group of points S_{1}, may be performed one or a predetermined number of times without performing the process of determining whether or not the input of corresponding points has been complete (step S**109**).

Additionally, the corresponding point acquirer **112** of the first embodiment may determines, when a new corresponding point is input through operations by the user, whether or not the input of the new corresponding point is to modify the already-input corresponding point, and delete the already-input corresponding point if it is determined that the input of the new corresponding point is to modify the already-input corresponding point. The determination of whether or not the input of the new corresponding point is to modify the already-input corresponding point is performed by, for example, using a predetermined threshold value. If the distance between the newly-input corresponding point and the already-input corresponding point is less than the predetermined threshold value, the corresponding point acquirer **112** determines that the input of the new corresponding point is to modify the already-input corresponding point, and deletes the already-input corresponding point closest to the newly-input corresponding point. Thus, by determining whether or not the input of the new corresponding point is to modify the already-input corresponding point, it is possible to easily modify the already-input corresponding point when a new corresponding point is input.

Further, the corresponding point acquirer **112** of the first embodiment may not only acquire a new corresponding point and modify the already-captured corresponding point, but also receive an instruction to delete a corresponding point. Moreover, the configuration may be such that when the corresponding point acquirer **112** deletes a corresponding point, the processes subsequent to step S**105** in the positioning process are performed again based on the remaining corresponding points (first and second groups of points S_{1 }and S_{2}).

**Second Embodiment**

**100**A according to a second embodiment. As shown in **100**A includes the perspective image generator **110**, the perspective image acquirer **111**, the corresponding point acquirer **112**, a first calculator **123**, the second calculator **114**, the display image generator **115**, and a first point group modifier **126**. The image processor **100**A of the second embodiment differs from the image processor **100** of the first embodiment (shown in **100**A includes the first calculator **123** in lieu of the first calculator **113** and further includes the first point group modifier **126**. Hereinafter, same reference numerals represent structural elements between the image processors **100** and **100**A, and explanations thereof are omitted.

Similar to the first calculator **113** of the first embodiment, the first calculator **123** of the second embodiment calculates a difference D in position of the target A based on the first and second groups of points S_{1 }and S_{z}, and the first to fourth parameters P_{1 }to P_{4}. Here, when there is a first group of points whose coordinates are modified by the first point group modifier **126**, the first calculator **123** of the second embodiment uses the first group of points modified and output from the first point group modifier **126**. Otherwise, the first calculator **123** uses the first group of points acquired by the corresponding point acquirer **112**.

The first point group modifier **126** receives the first group of points S_{1 }whose coordinates have been updated by the second calculator **114**, the second group of points S_{2 }acquired by the corresponding point acquirer **112**, the first and second perspective images PI_{1 }and PI_{2 }generated by the perspective image generator **110**, and the third and fourth perspective images PI_{3 }and PI_{4 }acquired by the perspective image acquirer **111**. The first point group modifier **126** modifies the updated coordinates of the first group of points S based on the first to fourth perspective images PI_{1 }to PI_{4 }and generates the modified coordinates of the first group of points S_{1}.

Hereinafter, the modifying process to be performed by the first point group modifier **126** is explained in detail. The first point group modifier **126** captures a partial image r^{(3) }having the predetermined size and a center at the n-th corresponding point (u_{n}^{(3)}, v_{n}^{(3)}) (n=1, . . . , N) on the third perspective image PI_{3}, which is included in the second group of points S_{2}. Here, N is the number of groups of corresponding points acquired by the corresponding point acquirer **112**. Additionally, the first point group modifier **126** captures a partial image r^{(1) }having a predetermined size and a center at the n-th corresponding point (u_{n}^{(1)}, v_{n}^{(1)}) (n=1, . . . , N) on the first perspective image PI_{1}, which is included in the first group of points S_{1}.

The first point group modifier **126** calculates a similarity C between the partial images r^{(1) }and r^{(3)}. The similarity is calculated by, for example, normalized cross correlation shown in the following Expression (21).

In Expression (21), C represents the similarity. As the similarity C becomes closer to 1, the similarity between the partial images r^{(1) }and r^{(3) }becomes higher. As the similarity C becomes closer to −1, the similarity between the partial images r^{(1) }and r^{(3) }becomes lower. I_{j}(u, v) represents a pixel value of the j-th perspective image PI_{j }(j=1, 2, 3, 4). u and v represent elements of a coordinate of a point included in an image region r^{(j)}. Ī_{j}^{r }represents an average of values of pixels included in the image region r(j) on the j-th perspective image PI_{j}.

Similarly, the first point group modifier **126** calculates a similarity C between partial images r^{(2) }and r^{(4)}. A partial image r^{(2) }has a predetermined size and a center at the n-th corresponding point (u_{11}^{(2)}, v_{11}^{(2)}) (n=1, . . . , N) on the second perspective image PI_{2}, which is included in the first group of points S_{1}. The partial image r^{(4) }has a predetermined size and a center at the n-th corresponding point (u_{n}^{(4)}, v_{n}^{(4)})(n=1, . . . , N) on the fourth perspective image PI_{4}, which is included in the second group of points S_{2}.

The first point group modifier **126** determines whether or not the similarity C calculated with respect to each corresponding point included in the first group of points S_{1 }is greater than or equal to a predetermined reference value. Then, the first point group modifier **126** determines to be a point to be modified, the corresponding point of the first group of points S_{T}, which has the similarity C that is smaller than the reference value.

The first point group modifier **126** captures, as a template, a partial image having a predetermined size and a center at the n-th corresponding point (u_{n}^{(3)}, v_{n}^{(3)}) on the third perspective image PI_{3}. The n-th corresponding point (u_{n}^{(3)}, v_{n}^{(3)}) on the third perspective image PI_{3 }corresponds to the n-th corresponding point (u_{n}^{(1)}, v_{n}^{(1)}) of the first group of points S_{1 }on the first perspective image PI_{1 }which is determined to be the point to be modified. The first point group modifier **126** detects a region on the first perspective image PI_{1}, which has the highest similarity to the captured template. The first point group modifier **126** replaces the n-th corresponding point (u_{n}^{(1)}, v_{n}^{(1)}) corresponding to the n-th corresponding point (u_{n}^{(3)}, v_{n}^{(3)}) with a center point of the region detected on the first perspective image PI_{1}, thus modifying the coordinates of the first group of points S_{1}. Here, the similarity to the template may be calculated by the normalized cross correlation or another method.

Thus, the first point group modifier **126** determines whether or not there is a corresponding point of the first group of points S_{1 }which requires modification. If there is a corresponding point of the first group of points S_{1 }which requires modification, the first point group modifier **126** modifies the coordinate of the corresponding point on the first or second perspective images PI_{1 }or PI_{2}. When the modification is made, the first point group modifier **126** supplies the modified first group of points S_{1 }to the first calculator **123**.

**100**A according to the second embodiment. The positioning process to be performed by the image processor **100**A of the second embodiment differs from the positioning process to be performed by the image processor **100** of the first embodiment (shown in **301** to S**303** are provided between steps **107** and S**108**. Hereinafter, same reference numerals represent same steps between the positioning processes of the first and second embodiments and explanations thereof are omitted.

After the second calculator **114** updates the coordinates of the first group of points S_{1 }in step S**107**, the first point group modifier **126** calculates the similarity C with respect to each corresponding point included in the updated first group of points S_{1 }(step S**301**).

The first point group modifier **126** compares to a reference value, the similarity C with respect to each corresponding point included in the first group of points S_{1}, thereby determining whether or not the corresponding point requires modification (step S**302**). If all the corresponding points require no modification (step S**302**: NO), the first point group modifier **126** proceeds to step S**108**.

If there is a corresponding point requiring modification (step S**302**: YES), the first point group modifier **126** performs the aforementioned modifying process on the corresponding point requiring modification, and supplies the modified first group of points S_{1 }to the first calculator **123** (step S**303**). Then, the first point group modifier **126** proceeds to step S**108**.

As explained above, according to the image processor **100**A of the second embodiment, the first point group modifier **126** determines whether or not there is a corresponding point of the first group of points S_{1 }which requires modification for such a reason that a corresponding point of the updated first group of points is deviated. This determination is made based on the similarity C between the partial images r^{(1) }and r^{(3) }or between the partial images r^{(2) }and r^{(4)}. When there is a corresponding point of the first group of points S_{1 }which requires modification, a coordinate of the corresponding point is modified. When deviation occurs to a corresponding point of the first group of points S_{1 }which is on the regenerated first or second perspective images PI_{1}′ or PI_{2}′, the coordinate of the deviated corresponding point is modified, thereby preventing deviation from occurring among the two corresponding points designated respectively on the first and second perspective images PI_{1 }and PI_{2}, and the two corresponding points designated respectively on the third and fourth perspective images PI_{3 }and PI_{4}. Thus, it is possible for the user to easily compare the first to fourth perspective images PI_{1 }to PI_{4}, and perform confirmation or modification of the corresponding points. The first point group modifier **126** is an example of a corresponding point modifier that calculates the above described similarity between the first and second partial images.

Here, the aforementioned possible structural modifications applicable to the image processor **100** of the first embodiment are also applicable to the image processor **100**A of the second embodiment.

**Third Embodiment**

**100**B according to a third embodiment. As shown in **100**B includes the perspective image generator **110**, the perspective image acquirer **111**, the corresponding point acquirer **112**, a first calculator **133**, the second calculator **114**, and the display image generator **115**. The image processor **100**B of the third embodiment differs from the image processor **100** of the first embodiment (shown in **100**B includes the first calculator **133** in lieu of the first calculator **113**. Hereinafter, same reference numerals represent same structural elements between the image processors **100** and **100**B, and explanations thereof are omitted.

The first calculator **133** receives the shifting amount of the first or second perspective image PI_{1 }or PI_{2 }based on an input received through operations by a user, the first and second parameters P_{1 }and P_{2}, and the position of the three-dimensional volume data Vd. The position of the three-dimensional volume data Vd is a three-dimensional coordinate of the target A indicated by the three-dimensional volume data Vd or a three-dimensional coordinate of a voxel of the voxel data. The position of the three-dimensional volume data Vd is data captured at the time the three-dimensional volume data Vd is captured. The shifting amount of the first or second perspective image PI_{1 }or PI_{2 }is input by, for example, the user operating the operation device **220** based on the first to fourth perspective images PI_{1 }to PI_{4 }displayed on the display system **200**.

_{1 }according to the third embodiment. As shown in _{1 }while comparing the first to fourth perspective images PI_{1 }to PI_{4 }displayed by the display system **200**. Thus, the shifting amount of the first perspective image PI_{1 }is input to the operation device **220**. In this case, the first calculator **133** receives the shifting amount of the first perspective image PI_{1 }received by the operation device **220**.

The first calculator **133** calculates a difference D in position of the target A between at the planning time and at the treatment time, based on the shifting amount of the first or second perspective image PI_{1 }or PI_{2 }input by the user. Hereinafter, a process for the first calculator **133** to calculate the difference D is explained in detail. _{1 }and PI_{2}, which are used for the first calculator **133** of the third embodiment to calculate the difference D.

In _{1 }represents a camera that captures a first perspective image PI_{1}. A virtual camera VC_{2 }represents a camera that captures a second perspective image PI_{2}. The position of the virtual camera VC_{1 }in the three-dimensional space is represented as X_{1}=(x_{1}, y_{1}, z_{1}). The position of the virtual camera VC_{2 }in the three-dimensional space is represented as X_{2}=(x_{2}, y_{2}, z_{2}). The position of the three-dimensional volume data Vd is represented as V=(V_{1}, V_{2}, V_{3}). The position of the three-dimensional volume data Vd is, for example, the position of the gravity center of the three-dimensional volume data Vd. The positions of the virtual camera VC_{1 }and the virtual camera VC_{2 }can be obtained from the first and second parameters P_{1 }and P_{2}.

The first calculator **133** calculates the three dimensional unit vectors v_{1}, v_{2}, and v_{3 }from the position X_{1 }of the virtual camera VC_{1}, the position X_{2 }of the virtual camera VC_{2}, and the position V of the three-dimensional volume data Vd. The unit vectors v_{1 }and v_{2 }are calculated from the following Expressions (22-1) and (22-2).

The unit vector v_{3 }can be calculated as a unit normal vector to a plane defined by the unit vectors v_{1 }and v_{2 }that extend from the position V of the three-dimensional volume data Vd. The shifting amount of the first or second perspective image PI_{1 }or PI_{2 }can be expressed by the three three-dimensional unit vectors v_{1}, v_{2}, and v_{3}. For example, when the first perspective image PI_{1 }is rotated by the angle θ around the center of the first perspective image PI_{1}, the three-dimensional volume data Vd rotates by the angle θ around the unit vector v_{1}. Using the Rodrigues's formula, a rotation matrix R′ can be calculated from the rotation angle around an arbitrary rotation axis. Additionally, when the first perspective image PI_{1 }is shifted by Δu in the horizontal direction of the first perspective image PI_{1 }and is shifted by Δv in the vertical direction thereof, that is, the first perspective image PI_{1 }is translated by the shifting amount (Δu, Δv), the translation vector t′ can be calculated as t′=Δu·v_{2}+Δv·v_{3}.

The rotation matrix R′ and the translation vector t′ represent the shifting of the position Y of the three-dimensional volume data Vd to the position X of the target A at the treatment time. For this reason, the rotation matrix R′ and the translation vector t′ can be converted into the rotation matrix R and the translation vector t which represent the difference D of the target A from the position X to the position Y. This conversion can be performed using the following Expressions (23-1) and (23-2).

*R=R′*^{T} (23-1)

*t=−R′*^{T}*t′* (23-2)

As explained above, the first calculator **133** of the third embodiment calculates the difference D of the target A (represented by the rotation matrix R and the translation vector t) based on the first or second perspective image PI_{1 }or PI_{2 }input through operations by the user. The image processor **100**B calculates the difference D of the target A based on the shifting amount of the image which is input by the user based on the user's experience, thereby making it possible to easily reduce the deviation between the first and third perspective images PI_{1 }and PI_{3 }and the deviation between the second and fourth perspective images PI_{2 }and PI_{4 }when the deviations are large. The first and second perspective images PI_{1}′ and PI_{2}′ regenerated after the deviations are reduced are compared to each other, thereby making it possible for the user to easily designate corresponding points.

Here, the aforementioned structural modification applicable to the image processor **100** of the first embodiment is also applicable to the image processor **100**B of the third embodiment.

Additionally, the image processor **100**B of the third embodiment may be configured such that the first calculator **133** is not provided, and the second calculator **114** receives the difference D input through operations by the user. In this case, the difference D is determined based on the experience of the user (such as a medical doctor or a surgeon), or selected by the user from multiple predetermined candidates for the difference.

According to the image processor and the treatment system including the image processor, which are of at least one of the above embodiments, the difference D that is the difference in position of the target A between at the planning time and at the treatment time is calculated. Then, the three-dimensional volume data Vd is shifted based on the difference D. Then, the first and second perspective images PI_{1 }and PI_{2}, which are to be compared with the third and fourth perspective images PI_{3 }and PI_{4 }captured at the treatment time, are generated from the shifted three-dimensional volume data Vd. Thus, it is possible to increase the similarity among the first to fourth perspective images PI_{1 }to PI_{4 }to be compared with one another by the user, thereby making it possible for the user to easily designate corresponding points.

The configurations of the image processors of any embodiments explained above may be combined. For example, in an initial state where setting of corresponding points is initiated, after the shifting amount is received through operations by the user and the difference D is calculated in the manner explained with respect to the image processor **100**B of the third embodiment, the subsequent processes may be performed in the manner explained with respect to the image processor **100** of the first embodiment or the image processor **100**B of the second embodiment. By this combination, when the difference D of the target A is large, precise calculation of the difference D can be performed after the deviation among the first to fourth perspective images PI_{1 }to PI_{4 }are reduced based on the user input, thereby enabling a reduction in time required for performing the positioning process.

Additionally, the descriptions have been given in each embodiment with respect to the configuration in which the image processor generates the first and second perspective images PI_{1 }and PI_{2 }from the three-dimensional volume data Vd shifted based on the difference D. However, the image processor may be configured to, instead of shifting the three-dimensional volume data Vd based on the difference D, shift, based on the difference D, the positions of the virtual cameras VC_{1 }and VC_{2 }respectively indicated by the first and second parameters P_{1 }and P_{2}. In other words, the image processor may be configured to change, based on the difference D, the virtually-viewing direction of the first perspective image PI_{1 }and the virtually-viewing direction of the second perspective image PI_{2}.

Further, each of the above image processors may be implemented by a general semiconductor integrated circuit, or a customizable electronic circuit such as an FPGA (field programmable gate array). Also each of the above image processors may be implemented by a processor and a memory that stores processor-executable instructions that, when executed by the processor, to cause the processor to perform the above-described acts or operations for the above-described image processing.

In other cases, one or more programs for the image processor of each of the above embodiments to perform the above-described acts or operations for the above-described image processing may be stored in a computer-readable recording medium, so that a computer system can read and execute the program stored in the computer-readable recording medium, thereby performing the positioning process.

Here, the “computer system” may include software such as an OS and one or more applications and hardware such as any peripheral device. Additionally, the “computer system” may include a WWW system that provides homepage providing or displaying environments.

Further, the “computer-readable storage medium” may include any non-transitory computer-readable storage mediums such as a storage device, such as a portable medium, for example, a flexible disk, a magneto optical disk, a ROM, or a CD-ROM, or a hard disk built in a computer system. Moreover, the computer-readable storage medium may also include a medium that temporarily stores a program, such as a volatile memory included in a computer system which serves as a server or client when the program is transmitted via a network such as the Internet, or a communication line such as a telephone line.

Additionally, the above program may be transmitted from a computer system storing that program in a storage device thereof to another computer system via a transmission medium or by carrier waves passing through a transmission medium. The “transmission medium” that transmits the program may include a medium having a function of transmitting information, such as a communication network, for example, the Internet, or a communication line, for example, a telephone line. Further, the above program may include a differential program in combination with the program already stored in the computer system.

While certain embodiments have been described, these embodiments have been presented by way of example only, and are not intended to limit the scope of the inventions. Indeed, the novel embodiments described herein may be embodied in a variety of other forms; furthermore, various omissions, substitutions and changes in the form of the embodiments described herein may be made without departing from the spirit of the inventions. The accompanying claims and their equivalents are intended to cover such forms or modifications as would fall within the scope and spirit of the inventions.

## Claims

1. An image processor comprising:

- a perspective image generator that generates a first perspective image of a target viewed in a first direction, based on three-dimensional volume data of the target;

- a perspective image acquirer that acquires a second perspective image of the target viewed in a second direction;

- a corresponding point acquirer that acquires one or more pairs of first and second corresponding points respectively on the first and second perspective images;

- a first calculator that updates the position of the one or more first corresponding points on the first perspective image, based on a difference in position between the first and second perspective images; and

- a display image generator that generates a first display image, the first display image including the first and second perspective images, the one or more first corresponding points updated on the first perspective image, and the one or more second corresponding points on the second perspective image,

- wherein the perspective image generator changes the first direction into a third direction based on the difference in position between the first and second perspective images, and generates an updated one, viewed in the third direction, of the first perspective image.

2. The processor according to claim 1, further comprising:

- a second calculator that calculates the difference by calculating a difference in position between the three-dimensional volume data and the target at the time the second perspective image is captured before the second perspective image is acquired by the perspective image acquirer, based on the one or more first and second corresponding points of the first and second perspective images.

3. The processor according to claim 2, wherein the corresponding point acquirer acquires one or more pairs of third and fourth corresponding points respectively on the first and second perspective images,

- the second calculator updates the difference in position between the first and second perspective images, based on the one or more pairs of first and second corresponding points and the one or more pairs of third and fourth corresponding points,

- the first calculator updates the positions of the one or more first and third corresponding points, based on the difference updated lastly in position between the first and second perspective images, and

- the display image generator generates an updated one of the first display image including the first perspective image and the one or more first and third corresponding points, which are updated lastly, and the second perspective image.

4. The processor according to claim 2, wherein the second calculator updates the difference, based on one or more undeleted pairs of the first and second corresponding points acquired, when any corresponding points have been deleted.

5. The processor according to claim 3, wherein the second calculator updates the difference, based on one or more undeleted pairs of the first and second corresponding points acquired, when any corresponding points have been deleted.

6. The processor according to claim 1, wherein the corresponding point acquirer determines whether a designated corresponding point on one of the first and second perspective images is a modified corresponding point to which the existent one of the one or more first and second corresponding points or an additional corresponding point to be added to the existent ones of the one or more first and second corresponding points, the corresponding point acquirer deletes a closest existent one, to the designated corresponding point, of the one or more first and second corresponding points in a case that the corresponding point acquirer determines that the designated corresponding point is the modified corresponding point.

7. The processor according to claim 1, further comprising:

- a corresponding point modifier that calculates a similarity between first and second partial images, the first partial image including the first corresponding point which position has been last updated by the first calculator, the second partial image including the second corresponding point which is paired with that first corresponding point,

- the corresponding point modifier detecting a region having a highest similarity to the second partial image from the first perspective image, if the similarity calculated is smaller than a predetermined reference value; and

- the corresponding point modifier modifying the position of the first corresponding point based on the region detected.

8. The processor according to claim 1, wherein the display image generator modifies, based on the difference, an epipolar line on the first perspective image, the epipolar line is associated with the one or more first corresponding points.

9. The processor according to claim 1, wherein

- the perspective image generator changes the third direction into a fourth direction based on the difference in position between the updated one, viewed in the third direction, of the first perspective image and the second perspective image, and

- the perspective image generator generates a further updated one, viewed in the fourth direction, of the first perspective image.

10. A treatment system comprising:

- an image processor;

- an irradiator that irradiates X-rays to a target;

- a detector that detects X-rays transmitted through the target and makes a second perspective image of the target viewed in a second direction based on the X-rays; and

- a display system,

- wherein the image processor comprises:

- a perspective image generator that generates a first perspective image of the target viewed in a first direction, based on three-dimensional volume data of the target;

- a perspective image acquirer that acquires the second perspective image from the detector;

- a corresponding point acquirer that acquires one or more pairs of first and second corresponding points respectively on the first and second perspective images;

- a first calculator that updates the position of the one or more first corresponding point on the first perspective image, based on a difference in position between the first and second perspective images; and

- a display image generator that generates a first display image, the first display image including the first and second perspective images, the one or more first corresponding points updated on the first perspective image, and the one or more second corresponding points on the second perspective image,

- wherein the perspective image generator changes the first direction into a third direction based on the difference in position between the first and second perspective images, and generates an updated one, viewed in the third direction, of the first perspective image, and

- wherein the display system displays the first display image.

11. The system according to claim 10, further comprising:

- a second calculator that calculates the difference by calculating a difference in position between the three-dimensional volume data and the target at the time the second perspective image is captured before the second perspective image is acquired by the perspective image acquirer, based on the one or more first and second corresponding points of the first and second perspective images.

12. The system according to claim 11, wherein the corresponding point acquirer acquires one or more pairs of third and fourth corresponding points respectively on the first and second perspective images,

- the second calculator updates the difference in position between the first and second perspective images, based on the one or more pairs of first and second corresponding points and the one or more pairs of third and fourth corresponding points,

- the first calculator updates the positions of the one or more first and third corresponding points, based on the difference updated lastly in position between the first and second perspective images, and

- the display image generator generates an updated one of the first display image including the first perspective image and the one or more first and third corresponding points, which are updated lastly, and the second perspective image.

13. The system according to claim 11, wherein the second calculator updates the difference, based on one or more undeleted pairs of the first and second corresponding points acquired, when any corresponding points have been deleted.

14. The system according to claim 12, wherein the second calculator updates the difference, based on one or more undeleted pairs of the first and second corresponding points acquired, when any corresponding points have been deleted.

15. The system according to claim 10, wherein the corresponding point acquirer determines whether a designated corresponding point on one of the first and second perspective images is a modified corresponding point to which the existent one of the one or more first and second corresponding points or an additional corresponding point to be added to the existent ones of the one or more first and second corresponding points, the corresponding point acquirer deletes a closest existent one, to the designated corresponding point, of the one or more first and second corresponding points in a case that the corresponding point acquirer determines that the designated corresponding point is the modified corresponding point.

16. The system according to claim 10, further comprising:

- a corresponding point modifier that calculates a similarity between first and second partial images, the first partial image including the first corresponding point which position has been last updated by the first calculator, the second partial image including the second corresponding point which is paired with that first corresponding point,

- the corresponding point modifier detecting a region having a highest similarity to the second partial image from the first perspective image, if the similarity calculated is smaller than a predetermined reference value; and

- the corresponding point modifier modifying the position of the first corresponding point based on the region detected.

17. The system according to claim 10, wherein the display image generator modifies, based on the difference, an epipolar line on the first perspective image, the epipolar line is associated with the one or more first corresponding points.

18. The system according to claim 10, wherein

- the perspective image generator changes the third direction into a fourth direction based on the difference in position between the updated one, viewed in the third direction, of the first perspective image and the second perspective image, and

- the perspective image generator generates a further updated one, viewed in the fourth direction, of the first perspective image.

19. An image processing method comprising:

- generating a first perspective image of a target viewed in a first direction, based on three-dimensional volume data of the target;

- acquiring a second perspective image of the target viewed in a second direction;

- acquiring one or more pairs of first and second corresponding points respectively on the first and second perspective images;

- updating the position of the one or more first corresponding points on the first perspective image, based on a difference in position between the first and second perspective images;

- generating a first display image, the first display image including the first and second perspective images, the one or more first corresponding points updated on the first perspective image, and the one or more second corresponding points on the second perspective image;

- changing the first direction into a third direction based on the difference in position between the first and second perspective images; and

- generating an updated one, viewed in the third direction, of the first perspective image.

20. The image processing method according to claim 19, further comprising:

- calculating the difference by calculating a difference in position between the three-dimensional volume data and the target at the time the second perspective image is captured before the second perspective image is acquired, based on the one or more first and second corresponding points of the first and second perspective images;

- acquiring one or more pairs of third and fourth corresponding points respectively on the first and second perspective images;

- updating the difference in position between the first and second perspective images, based on the one or more pairs of first and second corresponding points and the one or more pairs of third and fourth corresponding points;

- updating the positions of the one or more first and third corresponding points, based on the difference updated lastly in position between the first and second perspective images; and

- generating an updated one of the first display image including the first perspective image and the one or more first and third corresponding points, which are updated lastly, and the second perspective image.

**Patent History**

**Publication number**: 20150117605

**Type:**Application

**Filed**: Aug 15, 2014

**Publication Date**: Apr 30, 2015

**Patent Grant number**: 9533172

**Inventors**: Kyoka SUGIURA (Kawasaki), Ryusuke HIRAI (Tokyo), Yukinobu SAKATA (Kawasaki), Yasunori TAGUCHI (Kawasaki), Takeshi MITA (Yokohama)

**Application Number**: 14/460,402

**Classifications**

**Current U.S. Class**:

**Imaging (378/62);**Three-dimension (345/419)

**International Classification**: A61N 5/10 (20060101); G06T 7/00 (20060101); G06T 15/20 (20060101);